Modelling And Application Of China's Mid-and Long-term Coal Demand

In order to deal with the environmental problems caused by the current high-intensive exploitation and extensive use of coal resources,at the same time put forward policies and measures to ease the downward pressure on the coal market,it is very necessary to perform a scientific prediction on the trend and especially the peak of China's coal demand.First,based on the historical data of coal consumption and related factors such as GDP,coal price,industrial structure,total population,energy structure,energy efficiency,coal production and urbanization rate among 1987-2012,this study compared the prediction effects of five models.These models include vector autoregressive model(VAR),radial basis function(RBF)neural network model,genetic algorithm demand estimation model(GA-DEM),particle swarm optimization demand estimation model(PSO-DEM)and input-output model(IO).Through comparing the results of different models with the corresponding actual coal consumption,it is concluded that with a testing period of 2006-2012,PSO-DEM model have relatively optimal predicted effect for China's total coal demand,with MAPE(mean absolute percentage error)being close to or below 2%.The other models also have acceptable prediction effects,with MAPE all being below 5%.Second,based on the historical data of coal consumption and four primary factors(economic growth,energy structure,investment,and industrial structure)during the period 1981–2013,with particle swarm optimization algorithm as the core method,this study established two integrated models for coal demand prediction.The forecast results of the integrated models are robust.According to the results from the model with smaller mean absolute percentage error,combined with actual statistics,in the business as usual Economic New Normal scenario,China's coal demand peaked in 2013,and then a downward trend started with an average annual rate of decline of 4.06%.Through the common positive action of the main influential factors,the coal consumption and economic growth of China have been showing signs of decoupling.However,if the energy structure is slow to adjust,the peak year will be postponed,the peak value will obviously increase,and the rate of decrease after the peak will decrease.The effects of slower changes in investment in the coal industry are similar to the effects of slower changes in energy structure,except that the peak value is only slightly impacted.If changes in industrial structure are slow,the peak year and the peak value are barely impacted.Finally,a design scheme for the system of coal demand forecast was proposed in this study.First of all,the functional requirements and non-functional requirements were analyzed specifically in the design scheme.And then,the whole framework and specific implement process was conducted.What's more,the details about database design and database management were also involved in the following.